Gradient Descent Provably Optimizes Over-parameterized Neural Networks.
Simon S. DuXiyu ZhaiBarnabás PóczosAarti SinghPublished in: ICLR (Poster) (2019)
Keyphrases
- neural network
- learning rules
- back propagation
- pattern recognition
- cost function
- fuzzy logic
- artificial neural networks
- feed forward
- fault diagnosis
- fuzzy systems
- multilayer perceptron
- activation function
- objective function
- genetic algorithm
- loss function
- error function
- radial basis function
- neural nets
- feedforward neural networks
- database
- mumford shah functional
- training algorithm
- worst case
- decision trees
- information systems
- search engine
- data sets